High performance computing environments often require distributed high performance file systems. Such file systems may be responsible for storing and managing access to millions of files. In some cases, such file systems may have to provide files for thousands, or even millions of simultaneous users.
Supporting robust replication and restoration operations are important requirements for high performance file systems. Typically, distributed file systems may employ backup systems that mirror the data stored on the primary file system. However, for long running backup processes in operating in high performance computing environments backup and mirroring processed may be interrupted in the middle of the backup and/or restoration process. Such interruption may create inconsistent backups that may be difficult to or impossible to generate accurate point-in-time restorations. Thus, it is with respect to these considerations and others that the present invention has been made.